Microdialysis probe technology provides access to tissue interstitium for either sampling of diffusible tissue constituents or delivery of bioactive substances. For any analyte of interest, however, the relationship between the analyte concentration in the probe perfusate and in the tissue is a complex function of many factors, such as analyte molecular weight, analyte physicochemical properties, tissue properties, probe membrane properties, probe geometry, and perfusion rate. Better understanding of the relationship is needed in order to improve the quantitative usefulness of the technology. Current studies emphasize two factors in particular. One is the effect of trauma resulting from probe insertion into the tissue. The second is the strong influence of clearance processes in the tissue that remove analyte from the extracellular space, such as cellular uptake, chemical conversion and loss to blood through the microvasculature in the vicinity of the probe. Mathematical modeling incorporating these factors is used to describe diffusive and convective solute transport within the probe and in surrounding tissue. A principal outcome of the models is predictive expressions for the probe extraction efficiency. These modeling efforts have expanded the utility of calibration techniques for determining the extraction efficiency in vivo from perfusate concentration measurements. In addition to permitting estimation of analyte concentrations in tissue extracellular fluid, the magnitude of the extraction efficiency provides quantitative information about the tissue. These quantitative analyses require knowledge of physical properties of the probes, such as diffusive and convective permeabilities of probe membranes, that can be determined under well-characterized conditions in vitro. Applications of these quantitative approaches to microdialysis in the brain are being pursued in connection with alcoholism and studies of drugs of abuse. Other applications involve various normal tissues and tumors. Endogenous solutes of interest include neurotransmitters, particularly dopamine. Examples of exogenous substances employed are Zidovudine (AZT), cisplatin and analogs, fluconazole, ethanol, cocaine and opioids. Validation experiments in animals (mice, rats and primates) involve quantitative autoradiography, histology, and chemical assay of tissue surrounding the probe, as well as measurement of probe perfusate concentrations. [unreadable] [unreadable] In agreement with model predictions, we and others have previously shown that the extraction efficiency for dopamine in the brain decreases with reduction in the rate of extracellular clearance of this neurotransmitter. However, a recent study in the rat striatum suggested that the extraction efficiency may be insensitive to increases in dopamine clearance. This is not the case in mouse nucleus accumbens based on a study we completed this year. In mice treated with a long-acting kappa-opioid receptor antagonist, nor-binaltorphimine, the extraction efficiency was higher than in control animals. Model calculations indicated that the treatment increased the apparent rates of dopamine release and uptake approximately six-fold. [unreadable] [unreadable] Convective exchange of solutes between the perfusate and the tissue is usually assumed to be negligible in comparison to diffusive movement. However, a convective contribution may be unavoidable. In some cases convective enhancement may be desirable for augmenting the rate of solute delivery and the extent of tissue penetration. Our previous mathematical model of microdialysis has been expanded to incorporate convection either from the perfusate to the tissue or in the reverse direction. For analytes exhibiting concentration linearity, the revised model predicts that properly defined measures of mass- and concentration-based extraction efficiency are symmetric, i.e., they possess the same value whether the probe is sampling or delivering analyte. Without this symmetry, most probe calibration techniques would be invalid. Experiments are planned for testing this prediction.